Among other things, the cold war was an era of transition, ornamented with ambivalence, hypocrisy, inconsistency, and paradox. Bringing these issues into focus matters now because we arguably again inhabit an era of transition—from cold war to placeless conflicts, from nation-based structures to global flows, from American ascendance to American declension, from just plain weather to global warming, etc. Amid these puzzling shifts, scholars, policy makers, managers, and citizens are increasingly aware of, and uneasy about, our limited understanding of that sort of technology on which contemporary society relies to address complex problems, themselves often poorly specified. So perhaps revisiting an earlier technologically robust transition period, the cold war, focusing on military innovation and uncertainty, might provide a helpful perspective.

Realistically, there were two transitions in the half-century after 1939, one into and another out of the cold war. Neither was smooth, neither was a sharp and decisive event—the first Soviet atom test came well after severe East-West tensions had arced, whereas the collapse of the Soviet Union represented the final punctuation mark to a decade-long process of unraveling. Here I will focus on the “into the cold war” transition, with special attention to military-related technologies.

Though many compelling shifts entwined with the cold war’s onset, at least four are worth noting for our purposes: the establishment of a bipolar, global, politico-technical competition; the creation in the United States of a large, permanent standing army, fed by a restored draft; the parallel implantation of a permanent international intelligence arm of the executive branch; and the U.S. military’s increasing fascination with and embrace of technological innovations for warfare. These are all familiar to cold war scholars, but it is the fourth element that centrally animates this discussion. Until World War II and its aftermath, as Bart Hacker has argued, the U.S. Army and Navy were cautious to a fault in entertaining, testing, and introducing new technological elements into military operations. Few U.S. interwar military leaders welcomed new technologies, not least insofar as they unsettled basic operations and assumptions for army or navy practice. Regarding tanks, which Erich Ludendorff argued had been critical to Allied successes late in World War I, the U.S. Army stalled and stumbled, lagging the relatively few European innovations during these decades. Nonetheless, this conservatism collapsed by degrees during World War II. Why? The simplest explanation might center on competition. For twenty-three years after 1918, American military technologies were rarely and barely tested against “real time” competition. Starting in the later 1930s, military planners began realizing that thin funding of air, land, and seaborne technological development could be placing the United States several steps behind militant nations that had begun investing in advancing their technical capabilities. Blitzkrieg in Europe and the attack on Pearl Harbor confirmed this.

World War II altered the military’s technological mindset, given rapid design and implementation of radar and proximity fuses, innovations in code-breaking (particularly Enigma) and in communications, control, and instrumentation capabilities, along with the decisive feedback from military field operations that pushed redesign of aircraft, tanks, and special-purpose ships, and of course the Manhattan Project’s atomic bombs. Pursuing the competition motif, Russian T-34-85 tanks that overmatched Germany’s Tigers and the Nazis’ rocket-powered V-1 and V-2 flying bombs doubtless played a substantial role. This shift plunged military planners and procurement officers into increasingly difficult territory, the domain of technological uncertainty. The balkiness and erratic reliability of innovative and unproven technologies had offered persistent problems during the war. Breakdowns proved routine and often inexplicable; fixes and redesigns were necessary and of uncertain effect (the B-29 bombers’ piston engines were but one case in point). Often, weapons were continuously designed and reworked even as they were used operationally, with feedback loops from performance driving trial-and-error fixes, a practice termed concurrency. Yet when they performed as promised, such devices could deliver decisive advantages, altering the course of firefights and battles or even the trajectory of a war.

Technological uncertainty is the perennial companion of technological innovation, whether in wartime or peacetime, in the military or the private sector. As Bruno Latour long ago noted, when a new technical artifact or capability is emergent, even those who have designed and fabricated it cannot have an effective understanding of its capabilities, its operations in use, the materials of which it was made, or the science underlying its materiality. This is so precisely because it is rare in the history of technology and science that “knowledge why” precedes and shapes “knowledge how.” Indeed, humans have long made and made use of objects whose inner structures and relations were unknown. In many cases, exploring such phenomena has been irrelevant to the effectiveness of the object.

In other cases, as with steam engines and rails, field use generated breakdowns that had vast, costly implications, and therefore users developed protocols to track, classify, and learn from failures. This created routinized feedback cycles triggering periodic redesigns, which over time resulted in more stability, durability, and effectiveness for the technologies, in essence “freezing the design.” This built users’ confidence in reliability, simplified maintenance and repair, and set performance boundaries and expectations that reduced stochastic failures. Trial-and-error engineering improvements largely prevented boiler explosions and made rails less prone to cracking and more able to endure heavy traffic. Often, scientific understanding “caught up” with technological practice, further stabilizing the situation and the technology’s “platform.” In some cases, though, reliable performance, once in place, marginalized efforts to pursue a technology’s scientific foundations, perhaps until user attempts to stretch performance generated failures and thus triggered fresh initiatives toward understanding foundational conditions and relationships.

However, in the wake of World War II, a pair of decisive shifts appear to have dislocated this “normal technology” development pattern. The first shift during the cold war affected a number of industrial fields where urgent demand, funded by rival military establishments, propelled what I’m calling experimental development of highly complex, yet workable devices—despite insufficient usable or relevant science. Often these technologies’ operations were opaque to measurement and instrumentation, as with much jet engine technology, blocking the accumulation of knowledge. Often these devices involved engineering innovations that crossed multiple fields (metallurgy, fluid dynamics, combustion, etc.), with the result that failures were not instructive, but baffling. Second, a pattern of continuous innovation along many of these trajectories entailed that design changes multiplied and user expectations altered at rapid rates. This meant that no technologically stable platform could be realized so that iterations of use could squeeze out faults and allow remediation. In essence, continuous redesign in the context of incomplete (or underdeveloped) science created durable or, in Karl Weick’s terms, “permanent” technological uncertainties. Neither military nor commercial rivalry permitted a freezing of designs that in turn could allow learning from failures to generate deep knowledgeability and condition a technological stabilization, as seems to have happened so often in earlier generations. In consequence, stochastic failures followed redesigns in irregular order.

<>Continuous redesign and persistent technological uncertainty had multiple implications for U.S. cold war military contracting. Four will be sketched here. First, both cost expectations and planned schedules for development and delivery regularly proved unreliable. Contract cost overruns arose in these contexts as a matter of course, not as a sign of incompetence. Redesigns meant delays as well, and together time slippages and escalating costs compelled contract revisions and project reorientations. As one engineering designer mythically said to an air force general angered about delays: “If you want it bad, you’ll get it bad.” Repeated efforts to reconstruct procurement practices during the cold war could never master this primary dilemma: getting a novel device on time and on budget could easily mean getting a device that lacked innovation, was obsolescent at first use, worked unreliably, or all three.

Second, redesigns and technological uncertainty directly dislocated production dynamics, maintenance practices, and the management of logistics and supply. Orders for components had to be canceled or adjusted when parts failures or deficiencies forced redesigns. Production lot sizes shrank as flurries of changes had to be integrated into fabrication planning; hopes for mass production routinely faded if the artifacts’ designs could not be frozen, or if innovations elsewhere superseded capabilities stabilized technologies offered. As in World War II, morphing cold war military technologies were designated for assignment, maintenance, and repair in blocks, often using dash or version numbers or letters, as with combat aircraft, jet engines, and the M-series of army tanks and military vehicles. Matching parts to artifact variants and supplying them in optimal numbers presented major logistical and contracting challenges, complicated by the necessity of implementing updated maintenance bulletins for earlier models, which called for retrospective replacement of some redesigned components (frequently, engine components, fuel pumps, and batteries).

Third, redesigns and technological uncertainty affected operational deployment of military technologies in at least two ways. On the ground, users’ training and knowledgeability were persistently unstable. First, users discovered flaws and shortcomings in devices (and unanticipated capabilities, at times) by operating them in field conditions or by stressing them in ways not anticipated by designers and fabricators. Second, user learning could and did force changes in operational readiness (at times taking devices off-line). At the military staff level, these flows repeatedly generated organizational tensions and conflicts, most commonly between engineering and planning teams on one hand and operational leaders on the other. Planners and designers endorsed technological innovations that promised to deliver critical advantages in future combat, whether in an atomic war of annihilation or in what became the brushfire wars of containment. This meant embracing the complexities of experimental development, innovation, and technological uncertainty, much as did Admiral Hyman Rickover in a relentless drive to develop nuclear submarines and solid-fuel Polaris atomic missiles with underwater launch capacity.

By contrast, traditional force commanders sought functional, reliable, and durable weapons that could be deployed with confidence, under the general principle that it is soldiers, pilots, and marines who prevail in engagements, not technologies. Hence during the cold war’s first two decades, they often found continuous innovation more a problem than a solution and, perhaps most notably, long remained hostile to integrating computers into operations. Though it comes from aerospace manufacturing, a 1960 wall placard from McDonnell’s mockup of the initial Mercury capsule nicely captures the operational folks’ view: “When a change is not necessary, it is necessary not to change.” NASA saw most things from the engineering/planning perspective, so these organizational tensions ramified well beyond military circles, while arising amid the challenges of technological uncertainty.

The immense costs and considerable waste accompanying experimental development generated a fourth problem, external to technology and design—political uncertainty. This came in two principal flavors, military and budgetary. The military form centered on the internal politics of the services and rumors, ripples, or actual steps toward shifts in grand strategy. The prospect of nuclear war made fighter aircraft and main battle tank development obsolete. The air force’s prominence in missiles intensified rocket rivalries with its army parent and stimulated the triumphantly mobile Polaris innovations. Vietnam devalued ICBMs, as the prospect of global death receded slowly in the Lyndon Johnson years, but prioritized small arms and flying gunships (the AC series, with the AC-130 cycling quickly through ten versions). The Robert McNamara Defense Department sought to rein in costs associated with mission performance rivalries by pushing for crossover technologies that several services could use, but this seemed more to promote specification wars than effective development and acquisition.

The budgetary uncertainty variant flowed from administration and congressional anxiety that expanding missions and their associated technologies extended cost commitments laterally, while the raggedness of experimental development built each program’s expenditures vertically. As early as Dwight Eisenhower’s first term, special commissions began sounding alarms about soaring costs, but given that the executive chiefly tapped veteran administrators and managers for these evaluations, their reports unsurprisingly emphasized needs for organizational changes, contract reforms, and accounting controls. Fundamentally, it appears that policy makers never could quite get their arms around the difference between routine procurement of stable technologies and the profound irregularities central to creating innovative technologies, testing and redesigning them, encountering their recurrent “unknown unknowns,” and in time bringing them onstream as military assets.

Technological uncertainty disrupts planning and cost-estimating, dislocates production and maintenance, deranges operational practices, and triggers both political conflicts and utopian, dirigiste budgetary and contracting policies. In such contexts, then and now, ambivalence and equivocality can inform an effective stance toward decision making, as can a willingness to embrace inconsistencies, to value failures, and to shift our positions. Perhaps our practice of understanding innovation as incremental, logical, and recombinant, rather than as perilous attempts to step beyond the edges of the known and the reliable, has contributed to our difficulty in acknowledging these conditions for action in times of transition. The attempt to normalize and manage innovation frustrates the military still, and it surely has its corollaries in the private sector. My guess, though, is that the challenges of technological uncertainty will leave the stage when we choose to freeze all our designs and embrace the necessity “not to change.” Which is to say, not soon.

1 Management Learning 34 (2003): 379–82 (quote on 381).

Philip Scranton is University Board of Governors Professor, History of Industry and Technology, at Rutgers University.>/p>

Copyright© 2009, the Society for the History of Technology